Towards an understanding of peroxisome dynamics in mammalian cells

DES FACHBEREICHS MEDIZIN DER PHILIPPS-UNIVERSITÄT MARBURG

T

OWARDS AN UNDERSTANDING OF

P

EROXISOME

DYNAMICS IN MAMMALIAN CELLS

INAUGURAL-DISSERTATION ZUR ERLANGUNG DES DOKTORGRADES DER HUMANBIOLOGIE

(DR. RER. PHYSIOL.)

DEM FACHBEREICH MEDIZIN DER PHILIPPS-UNIVERSITÄT MARBURG

VORGELEGT VON

NINA ANNA MARIA BONEKAMP

Angenommen vom Fachbereich Medizin der Philipps-Universität Marburg am: 06.06.2012

Gedruckt mit Genehmigung des Fachbereichs.

Dekan: Prof. Dr. Matthias Rothmund Referent: Prof. Dr. Roland Lill Koreferent: Prof. Dr. Gerhard Schratt

1 INTRODUCTION ₁

1.1 Peroxisomes – an overview 1

1.1.1 Introducing a highly versatile organelle 1

1.1.2 Peroxisomal metabolism 3

1.1.3 Peroxisomes and reactive oxygen species 5

1.1.4 Peroxisomal disorders 7

1.1.5 Peroxisomal protein import 9

1.1.5.1 Import of peroxisomal matrix proteins 11 1.1.5.2 Insertion of peroxisomal membrane proteins 14

1.2 Peroxisome dynamics 16

1.2.1 Models of peroxisome biogenesis: “growth and division” vs. “de novo

synthesis” 16

1.2.2 The division machinery 18

1.2.2.1 The Pex11 family of proteins 19

1.2.2.2 Peroxisome fission 22

1.2.2.3 Recruiting DLP1 to peroxisomal membranes – Fis1, Mff 24

1.2.3 Peroxisome motility and inheritance 26

1.2.4 Regulation of peroxisome abundance 27

1.3 The peroxisome-mitochondria connection 30

1.4 Objectives 34

2 MATERIALS AND METHODS ₃₆

2.1 Equipment 36

2.2 Consumables 38

2.3 Chemicals and reagents 38

2.3.1 Chemicals 38

2.3.2 Loading dyes and markers 40

2.3.3 Kits 40

2.3.4 Cell culture reagents 41

2.4 Immunological reagents 41

2.4.1 Primary antibodies 41

2.4.2 Secondary antibodies 42

2.5 Molecular biology reagents 42

2.5.4.2 Internal primers 44

2.6 Frequently used buffers and solutions 44

2.7 Mammalian Cell lines 49

2.7.1 Cell lines 49

2.7.2 Stable cell lines 51

2.8 Cell culture 51

2.8.1 Cell passage 51

2.8.2 Generation of cell stocks 52

2.8.3 Stimulation of AR42J cells with dexamethasone 53

2.8.4 Mycoplasma detection 53

2.8.4.1 Hoechst staining 53

2.8.4.2 Mycoplasm detection by polymerase chain reaction (PCR) 54

2.8.5 Transient transfection of mammalian cells 55

2.8.5.1 Dietheylaminoethyl-(DEAE)-Dextran transfection 55

2.8.5.2 Electroporation 56

2.8.5.3 Lipofectamine 56

2.8.5.4 Polyethyleneimine-(PEI)-transfection 57

2.8.5.5 Turbofect 57

2.8.6 Generation of stable cell lines 58

2.8.7 Generation of hybridoma cells/ in vivo fusion assay 59

2.8.8 Stimulation of peroxisomal fusion with peroxisomal metabolites 60

2.8.9 Screen for inducers of peroxisome tubulation and/or proliferation 60

2.8.10 Induction of peroxisome proliferation 62

2.8.10.1 Bezafibrate 63

2.8.10.2 Eicosatetraynoic acid (ETYA) 63

2.8.11 Detection of ROS generation using 2′,7′-dichlorodihydrofluorescein

diacetate (2′,7′-dichlorofluorescein diacetate; H2DCFDA) 63

2.9 Microscopic techniques 64

2.9.1 Immunofluorescence 64

2.9.2 Lipid droplet staining 65

2.9.3 Epifluorescence microscopy 65

2.9.4 Confocal microscopy 66

2.9.5 Image deconvolution 66

2.9.6 Live cell imaging 67

2.9.7 Spinning disk confocal microscopy 68

2.9.8 Quantitative evaluation of peroxisome and mitochondrial dynamics 69

2.9.8.1 Evaluation of peroxisome and mitochondrial fusion in fixed cells 69

2.9.8.2 Analysis of peroxisomal interactions 69

2.9.8.3 Quantification of peroxisome morphology 70

2.10 Bioinformatic screening tools 70

2.10.2 In silico determination of potential phosphorylation sites 71

2.11 Biochemical methods 71

2.11.1 Preparation of peroxisome-enriched fractions 71

2.11.2 Preparation of cell lysates 72

2.11.3 Triton-X-100 (Tx100) extraction assay 73

2.11.4 Protein precipitation 73

2.11.4.1 Methanol-Chloroform precipitation (Wessel & Flugge, 1984) 73 2.11.4.2 Trichloroacetic acid (TCA) precipitation 73

2.11.5 Determination of protein concentration according to Bradford 74

2.11.6 One dimensional polyacrylamide gelelectrophoresis (SDS-PAGE) 74

2.11.7 Immunoblotting 76

2.11.8 Ponceau S staining 76

2.11.9 Enhanced chemiluminescence (ECL) for detection of proteins 76

2.11.10 Removal of antibodies from western blots (Stripping) 77

2.11.11 Immunoprecipitation (IP) 77

2.11.11.1 Radio-immunoprecipitation to validate cycloheximide efficiency 78 2.11.11.2 Radio-immunoprecipitation to determine protein phosphorylation (In

vivo phospholabelling) 79

2.11.12 Proteinase K digest 80

2.11.13 Carbonate treatment 81

2.12 Molecular biology techniques 82

2.12.1 Extraction of total RNA from mammalian cell lines 82

2.12.1.1 RNA-extraction using RNeasy Mini Kit (Qiagen) 82 2.12.1.2 RNA extraction using TriFast (PeqLab, (Chomczynski & Sacchi, 1987)) 83

2.12.2 Reverse Transcription (cDNA synthesis) 84

2.12.3 Polymerase chain reaction (PCR) 84

2.12.4 Semi-quantitative PCR (SQ-PCR) 86

2.12.5 Agarose gel electrophoresis 87

2.12.6 Restriction digest 88

2.12.7 DNA precipitation 88

2.12.8 Gel Extraction 89

2.12.9 Dephosphorylation of vector DNA 89

2.12.10 Ligation 89

2.12.11 Generation of chemically competent bacteria 90

2.12.12 Chemical transformation into competent bacteria 90

2.12.13 Plasmid preparation in a small (mini prep) and large scale (maxi

prep) 91

3.1.2 The occurrence of yellow peroxisomes points to peroxisomal fusion

events in CHO cells 96

3.1.3 Live cell imaging reveals close peroxisomal contacts and vivid

peroxisomal interactions without an exchange of matrix proteins 98

3.1.4 Transient peroxisomal interactions can potentially contribute to the

homogenization of the peroxisomal compartment 100

3.1.5 Transient peroxisomal interactions display a complex behaviour 103

3.1.6 Peroxisomal interactions: not be mistaken for fission 105

3.1.7 Peroxisomes do not exchange membrane proteins during peroxisome

interaction 106

3.1.8 Peroxisomal interactions do not increase after fatty acid or H2O2

treatment 108

3.1.9 Mitochondrial fusion proteins do not localize to peroxisomes 109

3.1.9.1 The outer membrane fusion proteins Mfn 1 and 2 do not localize to

peroxisomes 110

3.1.9.2 The inner mitochondrial membrane protein OPA1 is not targeted to

peroxisomes 112

3.1.10 Summary 113

3.2 Regulation of peroxisome dynamics: characterization of the peroxisomal membrane protein Pex11β and its N-terminal domain 114 3.2.1 Predicted positions of transmembrane domains within human

Pex11pβ 115

3.2.2 Pex11pβ is removed from the peroxisomal membrane by Triton-X-100

treatment after formaldehyde fixation 116

3.2.3 All human Pex11 isoforms behave like integral membrane proteins 118

3.2.4 Characterization of a newly available Pex11pβ antibody 119

3.2.5 Proteinase K digest of human Pex11pβ results in the formation of a 17

kD protease-protected fragment 121

3.2.6 In peroxisome-deficient cells, Pex11pβ is mistargeted to mitochondria 123

3.2.7 Upon mistargeting to mitochondria Pex11pβ retains its Tx100

sensitivity and orientation 125

3.2.8 Pex11pβ targeting to mitochondria depends on its N-terminal domain 127 3.2.9 Post-translational regulation of human Pex11pβ: In silico phospho

screening of mammalian Pex11β 128

3.2.10 Human Pex11pβ is not phosphorylated in COS-7 cells 130

3.2.11 Pex11pβ-mediated peroxisome membrane elongation is regulated by

homo-dimerization 131

3.2.12 Summary 133

3.3 Identification of novel stimuli altering peroxisome dynamics 134 3.3.1 6-hydroxydopamine induces DLP1-dependent fragmentation of

mitochondria and apoptosis in SH-SY5Y neuroblastoma cells but has

3.3.2 Alterations of peroxisome dynamics in response to oxidative stress 137 3.3.2.1 Overview of the model system employed to study alterations of

peroxisome dynamics in response to oxidative stress 137 3.3.2.2 Screening for alterations of peroxisomes in response to oxidative stress 138 3.3.2.3 Compartment-specific activation of KillerRed does not induce peroxules

or peroxisomal tubules 141

3.3.3 Dexamethasone treatment leads to an elongation of peroxisomes in

AR42J cells 144

3.3.4 One-time stimulation with dexamethasone is sufficient to induce

peroxisome elongation 147

3.3.5 The observed changes in peroxisome morphology after dexamethasone

treatment are reminiscent of the phenotype of Pex11pβ

overexpression 149

3.3.6 Pex11α and Pex11β are induced upon dexamethasone treatment 149

3.3.7 AR42J cells do not respond to bezafibrate with peroxisome

proliferation 153

3.3.8 Summary 154

4 DISCUSSION:TOWARDS AN UNDERSTANDING OF PEROXISOME DYNAMICS IN MAMMALIAN CELLS 156 4.1 Peroxisomal dynamics: Do mammalian peroxisomes fuse? 158

4.1.1 Unlike mitochondria, mature mammalian peroxisomes do not

exchange matrix or membrane components 158

4.1.2 Transient and complex peroxisomal interactions: a new dynamic

behaviour of mammalian peroxisomes 161

4.1.3 Peroxisomal versus mitochondrial dynamics 163

4.2 Regulation of peroxisome dynamics: Characterization of mammalian Pex11β

and its N-terminal domain 165

4.2.1 Human Pex11pβ – one integral membrane protein modulating the

morphology of two organelles 165

4.2.2 Regulation of human Pex11pβ by oligomerization, but not

phosphorylation 170

4.3 Novel stimuli altering peroxisome dynamics 174 4.3.1 Peroxisomes and neuronal apoptosis: no need for elongation at the

point of no return? 174

4.3.2 Stress-induced peroxisomal elongation: the nature of the signal is the

key 176

5 SUMMARY ₁₈₆ 6 ZUSAMMENFASSUNG 189 7 REFERENCES 193 8 APPENDIX ₂₃₀ 8.1 Supplementary Material 230 8.1.1 Supplementary Information 230

8.1.1.1 Theoretical Model explaining power law behaviour 230

8.1.1.2 Results Phospho-Screen HsPex11pβ 231

8.1.2 Supplementary Figures 232

8.1.3 Supplementary Movies 243

8.2 List of abbreviations 244

8.3 List of figures 248

8.4 Curriculum vitae 251

8.5 Verzeichnis der akademischen Lehrer 252

8.6 Acknowledgements 253

1

I

NTRODUCTION

1.1

PEROXISOMES – AN OVERVIEW

1.1.1

Introducing a highly versatile organelle

Unlike mitochondria or the Golgi apparatus, both of which were discovered in the very end of the 19th century, peroxisomes were only identified using electron microscopy in 1954 by Rhodin as a part of his PhD thesis and then termed microbodies (Rhodin, 1954) (Fig. 1.1). Originally regarded as a “fossil organelle” or the cell’s “garbage pail”, they have gained considerable interest upon their subsequent biochemical characterization by de Duve and Baudhuin who discovered that the peroxisomal matrix contains a high number of hydrogen peroxide (H2O2)-producing oxidases as well as catalase, an H2O2-degrading enzyme (De Duve & Baudhuin, 1966) (Fig. 1.1 B). This observation coined the more functional name “peroxisome” for the organelle.

Fig. 1.1: Ultrastructure of peroxisomes.

(A) Ultrathin sections of peroxisomes in the methylotrophic yeast Hansenula polymorpha after growth on methanol (Krikken et al., 2009). (B) Rat liver peroxisomes labelled with specific antibodies against catalase followed by protein A-gold staining (Fahimi, 1992). Note that gold particles representing catalase distribute over the entire matrix sparing the crystalline cores. (C) Peroxisomes in rat hepatoma cells stained by alkaline DAB cytochemistry (Schrader & Yoon, 2007). Note the close association of peroxisomes (black structures) and mitochondria (M) with the endoplasmic reticulum (ER). M, mitochondria; N, nucleus; P, peroxisome; V, vacuole. Bars, 0.5µm (A, C). Magnification (B) × 58,000.

The identification of catalase as a key peroxisomal enzyme led to the introduction of the alkaline 3, 3’-diaminobenzidine (DAB) reaction for catalase that enabled the morphological

the development of red blood cells and sperm (Luers et al., 2006) and are absent in

Apicomplexa phylum and amitochondriate parasites (Schluter et al., 2006).

In general, peroxisomes can be described as ubiquitous, single-membrane bound organelles that are devoid of DNA and contain a fine granular matrix (Fig. 1.1). About 85 genes in Homo

sapiens and 61 genes in Saccharomyces cerevisiae have been identified that encode

peroxisomal proteins, many of which are linked to peroxisome metabolism. Peroxisomes play a central role in lipid metabolism and detoxification, but also in the synthesis of ether phospholipids, bile acids and cholesterol (1.1.2), rendering them essential for cellular homeostasis and development. Their essential contribution to human health is exemplified by the severe consequences of peroxisomal dysfunctions (1.1.4). Despite this common ground, one of the most striking features of peroxisomes is their enormous plasticity, both on a metabolic and a morphological level. Peroxisomal enzyme composition varies immensely across species, even leading to the generation of specialized peroxisomal structures such as e.g. the glycosomes of trypanosomatids, the glyoxysomes of plants and the Woronin body of filamentous fungi, a very specialized peroxisome that serves the sole purpose of sealing septal pores upon injury (Jedd & Chua, 2000; Michels et al., 2005; Michels et al., 2006; Reumann & Weber, 2006). Interestingly, peroxisomal protein composition also varies within tissues and/or developmental stages of the same organism (Islinger et al., 2010), thus the term “multipurpose organelle” has proven to be a more than accurate designation for peroxisomes (Opperdoes, 1988; Islinger et al., 2010). Although peroxisomes display this striking heterogeneity in regard to their enzyme content, proteins involved in organelle biogenesis and maintenance – the so-called peroxins (Pex) - are evolutionary conserved throughout species. Pex genes have been identified based on genetic complementation in peroxisome-deficient yeast and mammalian cells (34 in yeast with approximately 20 mammalian and 23 plant homologues) (Kiel et al., 2006; Platta & Erdmann, 2007). Most of them are peroxisomal membrane proteins or associated with the membrane (1.1.5) (Table 1.2). Peroxisomes additionally show a remarkable plasticity in regard to their number and morphology. For instance, mammalian liver and kidney contain a high density of peroxisomes (accounting for around 2 % of the total hepatic protein content (Leighton et al., 1968)), whereas yeast cells in general only contain a few. Moreover, mammalian peroxisomes display a very heterogeneous morphology, since they may appear as spherical organelles with a diameter of around 0.1 – 0.2 µm, representing the common textbook image, but also enlarge to form rod-shaped (0.3 – 0.5 µm), long tubular structures (up to 5 µm) or even tubular-reticular networks (Yamamoto & Fahimi, 1987; Schrader et al., 1994; Schrader et al., 2000; Purdue & Lazarow, 2001). A

variety of peroxisomal shapes can also be observed on the ultrastructural level. Rat liver peroxisomes e.g. contain dense cores of urate oxidase, while bovine kidney peroxisomes display crystalline inclusions of α-hydroxyacid oxidase B, resulting in a rather polyhedric shape of the organelle (Hruban & Swift, 1964; Zaar et al., 1991). The Woronin bodies of

Neurospora crassa on the other hand appear as hexagonal crystals (Jedd & Chua, 2000).

Despite their enormous plasticity and dynamic behaviour, peroxisomes do not exist as isolated entities, but are intimately linked to other organelles such as lipid droplets, the ER and especially mitochondria (Schrader & Yoon, 2007; Camoes et al., 2009) (1.3) (Fig. 1.1 C). Peroxisome homeostasis in general needs to remain adaptable to the metabolic state of the cell which is ensured by a combination of peroxisome multiplication or proliferation, the removal of excess organelles by autophagy (pexophagy) as well as by processes of peroxisome inheritance and motility (1.2).

1.1.2

Peroxisomal metabolism

Peroxisomes are essential for human health and development, since they fulfil a variety of metabolic functions, both in the breakdown and synthesis of key substrates. Over 50 enzymes have been identified in the peroxisomal matrix that catalyze the β-oxidation of very long chain fatty acids (VLCFAs), prostaglandins and eicosanoids, the oxidation of D-amino acids, alcohols, polyamines and uric acid (in non-primates) as well as the α-oxidation of branched chain fatty acids. Furthermore, peroxisomes play a key role in the detoxification of xenobiotics, reactive oxygen species (ROS), glyoxylate and the biosynthesis of ether lipids, cholesterol and bile acids (van den Bosch et al., 1992; Ferdinandusse et al., 2002; Hogenboom et al., 2004; Kunze et al., 2006; van der Klei et al., 2006; Wanders & Waterham, 2006b) (Fig. 1.2). Several specialized functions have evolved with plants, yeast and protozoa generally displaying a broader spectrum of activities. These include an involvement in photorespiration, jasmonic acid and auxin synthesis in plants as well as penicillin synthesis in fungi (Tolbert, 1981; van den Bosch et al., 1992; Heupel & Heldt, 1994; Kiel et al., 2005b; Nyathi & Baker, 2006). The fact that e.g. the biosynthesis of penicillin in fungal peroxisomes can be environmentally induced renders peroxisome research an attractive field of interest for biotechnology (Kiel et al., 2005b). Besides the classical functions listed above, peroxisomes

metabolism. Detoxification of ROS is performed by catalase and a large set of resident peroxisomal antioxidant enzymes (1.1.3) (Fig. 1.2).

Fig. 1.2: Overview of the major peroxisomal metabolic pathways.

Peroxisomal metabolism in the mammalian liver (adapted from Baumgart et al., 1997; Schrader & Fahimi, 2008).

Peroxisomal lipid metabolism

While fatty acid β-oxidation occurs exclusively peroxisomal in yeasts and plants, it coexists with the mitochondrial fatty acid β-oxidation in animal cells (Cooper & Beevers, 1969; Kunau et al., 1988; Poirier et al., 2006; Shen & Burger, 2009). However, only the peroxisomal acyl-CoA oxidases (AOX) show substrate specificity for very long chain fatty acids (VLCFA, ≥ C24) and branched chain fatty acids (Mannaerts & Van Veldhoven, 1993; Mannaerts et al., 2000; Reddy & Hashimoto, 2001). Nonetheless, the basic set of fatty acid β-oxidation reactions carried out by both organelles remains the same and at the end of each cycle, fatty acids are shortened by two carbon atoms which are released as acetyl-CoA (Lazarow & De Duve, 1976; Wanders, 2004; Wanders & Waterham, 2006b) (Fig. 1.2). Upon reduction of fatty acid chain length to octanoyl-CoA, propionyl-CoA and acetyl-CoA, fatty acids are conjugated to carnitine and shuttled to mitochondria for full oxidation and energy production via the mitochondrial electron transfer chain (Ferdinandusse et al., 1999; Reddy & Hashimoto, 2001). Alternatively, they may serve as substrates for the generation of complex lipids within peroxisomes. Besides the prominent VLCFA docosahexaenoic acid (C26:0), peroxisomal substrates include pristanic acid, the bile acid intermediates di- and trihydroxycholestanoic acid (DHCA, THCA) and long-chain dicarboxylic acids. 3-methyl

oxidation machinery, hence they are degraded by peroxisomal α-oxidation, i.e. that they are oxidatively decarboxylated to generate 2-methyl-fatty acids which can be β-oxidized (Casteels et al., 2003; Jansen & Wanders, 2006; Wanders & Waterham, 2006a). In higher eukaryotes, peroxisomes are the sole site of α-oxidation.

Ether phospholipid synthesis

The term ether phospholipid denominates a special class of phospholipids which contain an ether-linkage instead of an ester linkage at the sn-1 position of glycerol. Ether lipids may contain an 1-O-alkyl linkage or an 1-O-alk-enyl linkage, the latter class is referred to as plasmalogens. Plasmalogens are important constituents of the neuronal myelin sheath, thus loss of peroxisomal functions is often accompanied by neurodegenerative processes (Faust et al., 2005; Hulshagen et al., 2008) (1.1.4). The biosynthesis of ether phospholipids involves the cooperation of peroxisomal and ER resident enzymes. Initial synthesis of acyl dihydroxyacetone phosphate (DHAP) and the generation of the eponymous ether bond is carried out by peroxisomal enzymes, while the generated alkyl-DHAP is then conversed into alkyl-glyceraldehyde 3-phosphate either in peroxisomes or at the ER. The final steps of ether phospholipid synthesis are performed exclusively at the ER (Heymans et al., 1983; Brites et al., 2004; Gorgas et al., 2006; Wanders & Waterham, 2006a).

Glyoxylate detoxification

Peroxisomes detoxify glyoxylate by the action of the enzyme alanine glyoxylate:aminotransferase (AGT) which catalyzes the transamination of glyoxylate to glycine with alanine. Unless it is detoxified, glyoxylate will be reduced to glycolate or oxidized to oxalate; in contrast to the water soluble glycolate, oxalate precipitates as calcium oxalate (Danpure, 2006).

1.1.3

Peroxisomes and reactive oxygen species

Although mitochondria are usually considered to be the major producer of ROS within the cell, peroxisomes contribute considerably to cellular ROS homeostasis due to their oxidative metabolism. In fact, they consume 20 % of the total amount of oxygen in rat liver while they

synthase (iNOS) that catalyzes the oxidation of L-arginine to nitric oxide (Stolz et al., 2002) (Table 1.1). In order to handle the massive generation of ROS, peroxisomes harbour not only catalase, their key enzyme, but a large set of other antioxidant enzymes, such as copper/zinc and manganese superoxide dismutase (SOD), glutathione peroxidase, peroxiredoxins and epoxide hydrolase (Dhaunsi et al., 1992; Singh et al., 1994; Singh, 1996; Immenschuh & Baumgart-Vogt, 2005; Bonekamp et al., 2009; Antonenkov et al., 2010; Bonekamp et al., 2011a; Fransen et al., 2011) (Table 1.1)

Table 1.1: Overview of ROS/RNS generated in mammalian peroxisomes (PO).

NOHLA, Nω-hydroxy-L-arginine, SOD, superoxide dismutase (adapted from Bonekamp et al., 2009)

Interesting findings regarding a dynamic, morphological response of peroxisomes to oxidative stress have been made in the plant system. Using live-cell imaging, plant peroxisomes were shown to respond to increasing oxidative stress in a two-fold manner: upon exposure to hydrogen peroxide or hydroxyl radicals, they were shown to develop small membrane protrusions – so-called peroxules - out of the previously spherical peroxisome in a time span of 10-120 seconds. Longer exposure or higher dosage of stressors resulted in a decline in peroxisome movement and their increased elongation and division (Sinclair et al., 2009). Similarly, mammalian peroxisomes respond to UV irradiation and H2O2 exposure with an elongation of the compartment (Schrader et al., 1999), hence peroxisomal elongation and/or proliferation might serve as a first line of defence counteracting oxidative stress. Changes in peroxisomal motility and a slight increase in number were also observed in plant cells after cadmium exposure (Rodriguez-Serrano et al., 2009). Notably, only the generation of intra-peroxisomal oxidative stress triggered a intra-peroxisomal response and the resulting motility change was assumed to improve inter-organellar cross-talk, uptake of metabolites or to scavenge ROS in places of need. Due to the highly oxidative environment, peroxisomes require several safeguarding mechanisms ensuring proper organelle function: catalase itself

activity in response to acute stress up to a certain threshold (Anand et al., 2009). Moreover, a peroxisomal Lon protease was identified in rat and H. polymorpha whose deletion led to an increase in protein oxidation, possibly removing damaged proteins to ensure functionality and thus contributing to intra-peroxisomal quality control (Kikuchi et al., 2004; Aksam et al., 2007). Apart from that, severely damaged organelles are removed by autophagy. Peroxisomes have further been linked to oxidative stress-related conditions such as neurodegeneration, carcinogenesis and aging (Cimini et al., 2009; Kou et al., 2011; Titorenko & Terlecky, 2011). In the case of aging, the fidelity of the peroxisomal matrix import system diminishes with age due to initial ROS-dependent modifications, resulting in a senescent cellular phenotype (Legakis et al., 2002; Terlecky et al., 2006). Moreover, the age-dependent “loss” of catalase leads to an accumulation of H2O2 and other ROS in peroxisomes which may facilitate H2O2 diffusion to the cytoplasm, where it may modulate signalling pathways and/or promote oxidative damage (Legakis et al., 2002). Aging also impairs the activities of other peroxisomal proteins, as was uncovered by mass-spectrometric analyses of kidney and liver peroxisomes (Mi et al., 2007).

1.1.4

Peroxisomal disorders

The pivotal role of peroxisomes in human health and development can be deduced from the severe phenotype of peroxisomal disorders, a group of inherited diseases in which either peroxisome biogenesis or single enzyme functions are disturbed. Based on this distinction, they are commonly subdivided into the peroxisome biogenesis disorders (PBDs) or peroxisomal (single) enzyme deficiencies (PEDs) (Steinberg et al., 2006; Wanders & Waterham, 2006a).

The PBDs are nowadays grouped into the Zellweger spectrum disorders (ZSD) and the rhizomelic chondrodysplasia punctata 1 (RCPD 1). ZSD is a collective term incorporating the Zellweger spectrum (ZS), neonatal adrenoleukodystrophy (NALD) and infantile Refsum disease (IRD), as patients suffer from the same clinical presentation of liver disease, neurodevelopmental delay, retinopathy and perceptive deafness within the first month of life, albeit to different degrees (Brosius & Gartner, 2002; Faust et al., 2005; Wanders & Waterham, 2005; Steinberg et al., 2006; Fidaleo, 2010). The ZS, initially referred to as

Mutations in at least 12 different peroxins were linked to the pathology of ZSD (Sacksteder & Gould, 2000; Steinberg et al., 2006). As Zellweger patients suffer from a complete absence of peroxisomes, toxic peroxisomal substrates (such as VLCFA, THCA and DHCA) accumulate while there is a shortage of essential peroxisomal products (such as plasmalogens) in every tissue. The consequences are especially detrimental in the brain, spinal cord and peripheral nerves in line with the central role of peroxisomes in the generation of plasmalogens which are essential for the formation of the myelin sheath (Faust et al., 2005; Hulshagen et al., 2008; Baes & Aubourg, 2009). As the accompanying developmental defects already start in utero, only supportive postnatal treatments are available based on dietary supplements and restrictions that merely counteract milder forms. The fact that ZSD presents with a low number of cases and a high variability restrains the development of new treatments. The other form of PBDs, RCPD1, presents differently from ZSD and is linked to mutations in PEX7, the gene encoding the import receptor for a subset of peroxisomal matrix proteins (Gould & Valle, 2000) (1.1.5.1).

The PEDs are grouped according to the biochemical pathway affected, e.g. ether lipid biosynthesis, peroxisomal β-oxidation, peroxisomal α-oxidation, glyoxylate detoxification and ROS metabolism (Wanders & Waterham, 2006a). The most common PED is the X-linked adrenoleukodystrophy (X-ALD) with an incidence of 1:15.000 males in France. Patients present with an accumulation of VLCFAs in the plasma, fibroblasts and other cell-types, resulting from mutations in the ABCD1 gene which encodes the adrenoleukodystrophy protein (ALDP). ALDP is an ABC transporter in the peroxisomal membrane, closely related to other metabolite transporters such as PMP70, and mediates the import of acyl-CoA esters of VLCFA into peroxisomes. Interestingly, its role was confirmed only recently (van Roermund et al., 2008; Wanders et al., 2010). In X-ALD, the accumulated VLCFAs were suggested to increase oxidative stress and thus lead to oxidative modification in nervous tissues (Schonfeld & Wojtczak, 2008). Furthermore, the VLCFA C26:0 may promote oxidative stress via mild inhibition of the mitochondrial electron transfer chain (ETC), which in combination with compromised cellular GSH levels, results in oxidative lesions (Fourcade et al., 2008). Other PEDs affecting peroxisomal β-oxidation were identified such as e.g. the AOX deficiency or D-bifunctional protein (DBP) deficiency. Interestingly, patient fibroblasts of AOX or DBP-deficient patients revealed peroxisomes that were enlarged in size and reduced in number (Chang et al., 1999). Except for X-ALD, all peroxisomal disorders are inherited in an autosomal-recessive manner. Similar to PBDs, PED treatment remains limited to supportive therapy based on the restriction of peroxisomal substrates or supplementation of

products. However, treatment options are available for patients suffering from primary oxaluria type I, a disease that results from AGT impairment. Interestingly, AGT mutations have been identified that introduce a mitochondrial targeting signal into the enzyme, leading to its mistargeting and subsequent loss of action (Danpure, 2006). As a result, oxalate deposits are formed in the kidneys (impairing renal function), but ultimately the failure in oxalate clearance results in its deposition in all tissues. As a treatment option, high oxalate concentrations can be diminished by inhibiting oxalate synthesis and increasing oxalate solubility (Danpure, 2005).

In addition to the classical peroxisomal disorders, other pathologies are now closer linked to peroxisomal function than previously assumed, including Alzheimer’s disease (AD), diabetes and cancer. Interestingly, peroxisome proliferation has neuroprotective effect counteracting β-amyloid (Abeta) toxicity (Santos et al., 2005). Thus, the modulation of peroxisomal and peroxisome-related proteins after acute and chronic insults with the toxic Abeta peptide was investigated to determine the neuroprotective role of peroxisomes upon Abeta-related oxidative injury (Cimini et al., 2009). Additionally, peroxisomal generation of H2O2 is involved in fatty acid-induced toxicity in insulin-producing pancreatic β-cells, thus contributing to the complex pathology of type 2 diabetes (Gehrmann et al., 2010; Elsner et al., 2011). Human carcinoma cells often display a significant reduction or even complete absence of peroxisomes (Lauer et al., 1999; Frederiks et al., 2010). These conditions might compromise cellular antioxidant capacity and facilitate further oxidative DNA damage, thus contributing to a more malignant behaviour. Furthermore, disorders affecting proteins of the peroxisomal growth and division machinery (the dynamin-like protein 1), and thus peroxisome dynamics, have been identified and are addressed in detail later on (Waterham et al., 2007) (1.2.2.2).

1.1.5

Peroxisomal protein import

As peroxisomes are devoid of DNA, and thus all peroxisomal proteins are encoded in the nucleus, matrix and membrane proteins are synthesized on free ribosomes in the cytosol and imported post-translationally into pre-existing organelles (Lazarow & Fujiki, 1985). The processes of peroxisomal matrix and membrane protein import are mediated by independent

been characterized (Kiel et al., 2006; Platta & Erdmann, 2007). An overview is given in Table 1.2.

Peroxin Organism Localization Domains Proposed function

Pex1p m p f y membrane

(cytosol)

AAA ATPase Matrix protein import, export of Pex5p

Pex2p m p f y integral PMP RING finger Matrix protein import, translocation

Pex3p m p f y integral PMP Membrane biogenesis,

PMP import

Pex4p p f y peripheral

PMP

E2 enzyme Matrix protein import, Pex5p ubiquitination

Pex5p m p f y Cytosol/

membrane

TPRs Matrix protein import, PTS1 (and PTS2) receptor

Pex6p m p f y membrane

(cytosol)

AAA ATPase Matrix protein import, export of Pex5p

Pex7p m p f y Cytosol/

membrane

WD40repeats Matrix protein import, PTS2 receptor

Pex8p f y peripheral

PMP (matrix)

Matrix protein import

Pex9p Yl (ORF wrongly identified, antisense sequence of Pex26p)

Pex10p m p f y integral PMP RING finger Matrix protein import, translocation

Pex11p m p f y (integral) PMP Proliferation and

division

Pex12p m p f y integral PMP RING finger Matrix protein import translocation

Pex13p m p f y integral PMP SH3 Matrix protein import,

docking

Pex14p m p f y (integral) PMP Coiled-coil Matrix protein import, docking

Pex15p Sc integral PMP Matrix protein import,

Pex1p/Pex6p anchor

Pex16p m p f Yl (integral) PMP Membrane biogenesis

Pex17p y peripheral

PMP

Coiled-coil Matrix protein import, docking

Pex18p Sc Cytosol/

membrane

Matrix protein import, PTS2 import Pex19p m p f y Cytosol/ membrane Farnesylation motif Membrane biogenesis, PMP import Pex20p f y Cytosol/ membrane

Matrix protein import, PTS2 import

Pex21p Sc Cytosol/

membrane

Matrix protein import, PTS2 import

Pex22p p f y integral PMP Matrix protein import,

Pex4p anchor

Pex23p f y integral PMP Dysferlin Proliferation

Pex25p y peripheral PMP

Proliferation

Pex26p m f integral PMP Matrix protein import,

Pex1p/Pex6p anchor

Pex27p Sc peripheral

PMP

Proliferation

Pex28p Sc integral PMP Proliferation

(Pex24p ortholog)

Pex29p y integral PMP Proliferation

Pex30p Sc integral PMP Dysferlin Proliferation

(Pex23p ortholog)

Pex31p Sc integral PMP Dysferlin Proliferation

Pex32p y integral PMP Dysferlin Proliferation

Pex33p Nc membrane Coiled-coil Matrix protein import,

Biogenesis

Pex34p y integral PMP Proliferation

Table 1.2: Overview of the peroxins (Pex).

Organisms: m, mammals; p, plants; f, filamentous fungi; y, yeasts; Nc, N. crassa; Sc, S. cerevisiae; Yl, Y.

lipolytica. RING, really interesting new gene; SH3, Src-Homology 3.

1.1.5.1 Import of peroxisomal matrix proteins

Peroxisomal matrix protein import is accomplished using four consecutive steps: the binding of the cargo protein to its import receptor, the docking of the receptor-cargo complex to the peroxisomal membrane, membrane translocation of the cargo and receptor recycling (for review, see Ma & Subramani, 2009; Lanyon-Hogg et al., 2010; Rucktaschel et al., 2011) (Fig. 1.3).

Depending on their inherent peroxisomal targeting signal (PTS), peroxisomal matrix proteins utilize either the PTS1-Pex5p or the PTS2-Pex7p-mediated import pathway. The majority of peroxisomal proteins contain a PTS1 at their extreme C-terminus. It was initially identified as the tri-peptide sequence SKL in firefly luciferase (Gould et al., 1987), but has been expanded to the general consensus sequence [S/A/C]-[K/R/H]-[L/M] (Gould et al., 1989; Aitchison et al., 1991; Elgersma et al., 1996; Purdue & Lazarow, 1996; Kragler et al., 1998; Lametschwandtner et al., 1998). Some proteins (e.g. mammalian catalase) require additional interactions or upstream sequences to enhance Pex5p-binding specificity (Maynard & Berg, 2007; Ma & Subramani, 2009). Upon completion of protein synthesis, the PTS1 is recognized by its specific import receptor, the cytosolic Pex5p (Brocard et al., 1994; Terlecky et al.,

repeats and the PTS1 peptide. There is an ongoing discussion about the stochiometry of Pex5p. It was shown to act as a monomer in solution (Costa-Rodrigues et al., 2004; Shiozawa et al., 2009), but also dimeric or tetrameric structures have been reported (Madrid et al., 2004). The second peroxisomal targeting signal, the N-terminal PTS2, was first identified in rat liver thiolase and is a conserved nonapeptide with the consensus sequence [R/K]-[L/V/I]-X5-[H/Q]-[L/A] (Swinkels et al., 1991; Rachubinski & Subramani, 1995). Notably, its presence varies greatly across species. While the PTS2-Pex7p-mediated import pathway is completely absent in the nematode C. elegans (Motley et al., 2000), it is only maintained in the yeast S. cerevisiae for the import of 3-ketoacyl thiolase and Gpd1 (Grunau et al., 2009; Jung et al., 2010), whereas about 60 proteins (a third of all matrix proteins) contain a PTS2 in

A. thaliana (Reumann et al., 2009). Unlike the PTS1, the PTS2 is cleaved from the nascent

protein upon import (Helm et al., 2007; Kurochkin et al., 2007). PTS2 proteins are recognized by their import receptor Pex7p, a member of the WD40 class of proteins. Although Pex7p itself appears to act as a monomer, it requires the assistance of auxiliary co-receptors (Rehling et al., 1996; Mukai & Fujiki, 2006; Grunau et al., 2009; Lanyon-Hogg et al., 2010). Pex5pL functions as a co-receptor for Pex7p in mammals and plants (Otera et al., 2000), while Pex18p/Pex21p (S. cerevisiae) or Pex20p (H. polymorpha, N. crassa, P. pastoris and Y.

lipolytica) are utilized in different yeast species (Einwachter et al., 2001; Otzen et al., 2005;

Leon et al., 2006b). These co-factors share structural similarities (Dodt et al., 2001; Schliebs & Kunau, 2006). Other proteins lacking a PTS such as S. cerevisiae acyl-CoA oxidase or mammalian Cu/ZnSOD enter peroxisomes using other pathways, for example by employing a piggy back mechanism via association with PTS1 proteins (McNew & Goodman, 1994; Yang et al., 2001; van der Klei & Veenhuis, 2006; Islinger et al., 2009).

Fig. 1.3: Peroxisomal matrix protein import.

The receptor cycle in the yeast S.

cerevisiae (see text for details) (adapted

from Rucktaschel et al., 2011). Note that Pex17p and Pex8p do not exist in mammals. Moreover, Pex18p/Pex21p action is replaced by Pex5pL and the function of Pex22p/Pex4p is fulfilled by UbcH5a/b/c. In mammals, Pex26p anchors Pex6p at the membrane instead of Pex15p.

After cargo binding, the cargo-receptor complex docks at the peroxisomal membrane upon interaction with the resident docking complex, composed of the proteins Pex13p, Pex14p and additionally Pex17p in yeast. Importantly, neither Pex13p nor Pex14p have the capability to bind cargo directly, thus Pex5p needs to stay complexed with its cargo during the import procedure. Membrane translocation of peroxisomal matrix proteins is thought to occuran extended shuttle mechanism that allows the receptor-cargo complex to pass completely the membrane (Kunau, 2001). In line with this, Pex5p, Pex7p and Pex20p are cycling receptors that always remain protease-protected to a certain degree (Gouveia et al., 2003; Leon et al., 2006a; Leon et al., 2006b). Notably, peroxisomes possess the striking ability to import fully folded and even oligomeric proteins (Glover et al., 1994; McNew & Goodman, 1994). Furthermore, even gold particles coated with a PTS1 of 9 nm size are imported (Walton et al., 1995), which is achieved without disruption of membrane compartmentalisation or any indication of the existence of static membrane pores (analogous to nuclear pores). To accommodate these features, a transient membrane pore composed of Pex5p oligomers has been proposed to form at the peroxisomal membrane (Erdmann & Schliebs, 2005) for whose existence evidence has been provided recently (Meinecke et al., 2010). Nevertheless, another elegant model was proposed which suggests that peroxisomal cargo can reach the peroxisomal membrane enfolded by Pex5p due to its natively unfolded state and might thus be translocated due to membrane embedding by Pex5p monomers (Grou et al., 2009a). The membrane translocation step itself is independent of ATP (Gouveia et al., 2003; Miyata & Fujiki, 2005). Cargo release into the peroxisomal lumen was suggested to be mediated by Pex8p, a protein which contains both PTS and therefore competes with Pex5p and Pex7p; however it has only been identified in yeast and fungi so far (Rehling et al., 2000). Cargo receptor recycling is an ATP-dependent process that requires the action of two protein subcomplexes and includes mono-ubiquitination of Pex5p. The so-called receptor-release complex consists of Pex1p and Pex6p, two proteins of the AAA family of ATPases, and additionally Pex4p and Pex22p in yeast (Ma & Subramani, 2009; Rucktaschel et al., 2011). Pex1p interacts with Pex6p which is anchored at the peroxisomal membrane by Pex15p (S.

cerevisiae) or Pex26p (mammals) (Matsumoto et al., 2003). The RING-finger proteins

(Pex2p, Pex10p and Pex12p) are E3 protein-ubiquitin ligases, while Pex22p serves as a docking site for the E2 ubiquitin-conjugating enzyme Pex4p (Wiebel & Kunau, 1992; Platta

Grou et al., 2009b) upon which it is extracted from the peroxisomal membrane by the action of Pex1p and Pex6p (Miyata & Fujiki, 2005; Platta et al., 2005). Subsequent de-ubiquitination of Pex5p can either occur by de-ubiquitinating enzymes or non-enzymatic processes (Grou et al., 2009b). If there is a delay in receptor release, Pex5p is poly-ubiquitinated, resulting in its degradation by the 26S proteasome (Kiel et al., 2005a; Platta et al., 2007) which provides a quality control mechanism for peroxisomal protein import. Poly-ubiquitination of Pex5p is thought to occur by the action of Pex2p and Pex10p (Williams et al., 2008; Platta et al., 2009). Ubiquitination of the PTS2 co-receptors Pex18p and Pex20p was also observed (Brown & Baker, 2008).

1.1.5.2 Insertion of peroxisomal membrane proteins

If peroxisomal matrix protein import is impaired, cells still contain peroxisomal remnant structures, the so-called ghosts, which retain a full set of peroxisomal membrane proteins (PMPs) (Santos et al., 1988; Brown & Baker, 2003; Schrader & Fahimi, 2008). Thus, peroxisomal membrane protein insertion is mediated in a manner completely distinct from matrix protein import (Fig. 1.4). The complete absence of peroxisomal membranes is only observed upon deletion or impairment of the peroxins Pex3p, Pex16p and Pex19p (Hohfeld et al., 1991; Baerends et al., 1996; Eitzen et al., 1997; Honsho et al., 1998; Matsuzono et al., 1999), identifying those as key factors of peroxisome membrane biogenesis. In contrast to the well-defined peroxisomal matrix targeting signals, no clear consensus sequence was identified for PMPs. However, specific membrane targeting sequences (mPTS) were determined in a variety of proteins displaying a different topology (Van Ael & Fransen, 2006). Depending on the protein, the mPTS varies greatly in length, but is usually comprised of a transmembrane segment and a cluster of basic amino acids (Baerends et al., 2000; Honsho & Fujiki, 2001; Jones et al., 2001; Rottensteiner et al., 2004). Peripheral peroxisomal membrane proteins harbour an mPTS and a protein interaction domain (Girzalsky et al., 2006). Most PMPs are targeted to the peroxisomal membrane in a manner dependent on Pex19p interaction (Class I PMPs) (Sacksteder et al., 2000; Snyder et al., 2000; Fransen et al., 2001; Halbach et al., 2005; Hadden et al., 2006; Halbach et al., 2006). Pex19p is farnesylated and shuttles between the cytosol and the peroxisomal membrane (Kammerer et al., 1997; Gotte et al., 1998; Rucktaschel et al., 2009). The multifunctional Pex19p simultaneously serves as an import receptor as well as a chaperone for newly-synthesized PMPs and has further been characterized as both an insertion factor and assembly/disassembly factor at the peroxisomal membrane (Fransen et al., 2001; Fransen et al., 2004; Jones et al., 2004; Shibata et al., 2004).

Shibata et al., 2004; Fransen et al., 2005; Matsuzono et al., 2006), the N-terminal part mediates peroxisomal membrane targeting through interaction with Pex3p.

Fig. 1.4: Peroxisomal membrane protein insertion. (upper panel) Topogenesis of peroxisomal membrane proteins. Two routes are proposed for the targeting of peroxisomal membrane proteins (PMPs). Class I proteins are directly imported into existing peroxisomes. Class II proteins are first targeted to ER where they concentrate in pre-peroxisomal vesicles which then are targeted to existing peroxisomes or function as an origin for de novo formation of peroxisomes.

(lower panel) Pex19p-dependent import of PMPs. Class I peroxisomal membrane proteins (PMPs) harbour a peroxisomal membrane protein targeting signal (mPTS) which is recognized in the cytosol by the import receptor and/or PMP-specific chaperone Pex19p. Cargo-loaded Pex19p docks to the peroxisomal membrane via association with its docking factor Pex3p. Then the PMP is inserted into the membrane in an unknown manner but presumably with assistance of Pex19p, Pex3p and, in some organisms, Pex16p (adapted from Rucktaschel et al., 2011).

Upon binding of Pex19p to newly-synthesized PMPs, the cargo-receptor complex is recruited to the peroxisomal membrane via the docking factor Pex3p, an integral membrane protein (Fang et al., 2004; Fransen et al., 2005; Matsuzono et al., 2006). It binds the cargo-receptor complex with a higher affinity than Pex19p alone which thus facilitates docking of the complex to the peroxisomal membrane (Soukupova et al., 1999; Ghaedi et al., 2000a; Ghaedi et al., 2000b; Hunt & Trelease, 2004; Haan et al., 2006; Pinto et al., 2006). In addition to that, Pex3p was implied to be involved in peroxisome inheritance in S. cerevisiae (Chang et al., 2009; Munck et al., 2009). Subsequently, the PMP is inserted into the peroxisomal membrane by an unknown mechanism and Pex19p is recycled back to the cytosol (Matsuzono & Fujiki, 2006).

1997; Honsho et al., 1998; Honsho et al., 2002) where it was suggested to act as an intra-peroxisomal regulator of organelle fission (Eitzen et al., 1997). Interestingly, mammalian Pex16p was shown to be co-translationally inserted into the ER prior to its trafficking to peroxisomes (Kim et al., 2006). In line with this, a small number of PMPs, such as Pex3p and Pex16p, was suggested to be inserted into peroxisomes in a Pex19p-independent manner (Class II PMPs) due to their lack of a Pex19p-binding domain (Fujiki et al., 2006). They are initially targeted to the ER, but, in order to release e.g. Pex3p from the ER, Pex19p is required (Hoepfner et al., 2005; Kragt et al., 2005; Tam et al., 2005). However, full-length Pex3p was also shown to directly interact with Pex19p and then be inserted into the peroxisomal membrane in a Pex16p-dependent fashion (Matsuzaki & Fujiki, 2008). Therefore, peroxisome membrane biogenesis can be initiated by the formation of a pre-peroxisomal membrane that carries either Pex3p or Pex16p, both of which are capable to develop into mature peroxisomes. Depending on the peroxin present (Pex3p or Pex16p), the respective missing PMP is targeted and inserted in a Pex19p-dependent manner to generate import competent peroxisomes.

1.2

PEROXISOME DYNAMICS

1.2.1

Models of peroxisome biogenesis: “growth and division” vs. “de novo

synthesis”

Peroxisomes were once implied to originate from virtually any organelle, including the Golgi complex, lysosomes and the ER (Novikoff & Essner, 1960; Rouiller & Jezequel, 1963; Novikoff & Shin, 1964). Subsequently, peroxisomal proteins were shown to be synthesized on free ribosomes in the cytosol (Goldman & Blobel, 1978) and then imported post-translationally into peroxisomes, therefore, the classical “growth and division” model was proposed (Lazarow & Fujiki, 1985). According to the latter, peroxisomes grow by import of protein and lipid components into pre-existing peroxisomes and their subsequent division into smaller organelles, rendering them autonomous like mitochondria. The protein machinery orchestrating peroxisomal growth and division will be discussed in detail in the next section (1.2.2).

In the last decade, strong evidence was provided for an alternative mechanism of peroxisomal

Titorenko & Mullen, 2006; Hettema & Motley, 2009; Saraya et al., 2010; Nuttall et al., 2011). A close relationship between the ER and peroxisomes has long been indicated; e.g. ultrastructural studies demonstrated close contacts between peroxisomes and the ER (Novikoff & Novikoff, 1972), including direct, luminal connections between the organelles in mouse dendritic cells (Geuze et al., 2003). Furthermore, some peroxisomal proteins such as Pex16p are known to travel to the ER prior to peroxisomes in Y. lipolytica, mammals and plants (Titorenko & Rachubinski, 1998; Karnik & Trelease, 2005; Kim et al., 2006). The exciting observation that yeast or mammalian cells devoid of peroxisomes due to deletions of Pex3p, Pex16p or Pex19p (1.1.5.2) are able to form peroxisomes de novo from the ER upon re-introduction of the respective missing gene finally challenged the prevailing model of an autonomous peroxisomal growth and division {Faber, 2002, Haan, 2006`, Hoepfner`, 2005`, Kim`, 2006`, Kragt`, 2005`, Matzuzuno`, 1999`, Muntau`, 2000`, South`, 1999`, Titorenko`, 2001}. In different yeast species, de novo formation of peroxisomes was initiated by re-introduction of Pex3p which localized to specific spots at the ER membrane where it budded off in a Pex19p-dependent manner to generate a pre-peroxisomal structure (Hoepfner et al., 2005; Tam et al., 2005; Haan et al., 2006; Titorenko & Mullen, 2006). After subsequent assembly of the membrane and matrix protein import machinery, pre-peroxisomes acquire import-competence. In regard to the composition of pre-peroxisomal vesicles, at least de novo formation itself is independent of COPI- or COPII-dependent transport (Matsuzono et al., 1999; South et al., 2000; Voorn-Brouwer et al., 2001; Kim et al., 2006), although the COPII vesicle component Sec16B was recently shown to be involved in the ER export of Pex16p and Pex3p (Yonekawa et al., 2011). Peroxisome biogenesis from the ER is further suggested to be facilitated by the ER components Sec20, Sec39 and Dsl1 as well as Pex19p (and yet unidentified components) in S. cerevisiae (Perry & Rachubinski, 2007; Perry et al., 2009; Agrawal et al., 2011; Lam et al., 2011) and Emp24, Pex25p as well as Rho1 in H. polymorpha (Saraya et al., 2011). The contribution of the ER to PMP trafficking or peroxisomal de novo formation under wild-type conditions remains controversial. Elegant pulse-chase experiments in S. cerevisiae indicated that peroxisomes divide by growth and division under wild-type conditions, but that loss of peroxisomes can then be compensated by a slower, de novo pathway (Motley & Hettema, 2007). However, others suggest a major contribution of the de

of components between the two organelles might be facilitated via contact sites or direct luminal connections, as peroxisomes were observed in close apposition with the ER (Geuze et al., 2003 2003, Tabak, 2008). In line with this, a non-vesicular transfer of phospholipids between the ER and peroxisomes was shown to be completely independent of Pex3p or any Sec components (Raychaudhuri & Prinz, 2008).

Fig. 1.5: Schematic view of peroxisome dynamics and interactions in mammalian cells.

Most peroxisomal matrix and membrane proteins (Class I PMPs) are synthesized on free polyribosomes in the cytosol and are post-translationally imported into pre-existing organelles. Other membrane proteins (Class II PMPs, early peroxins; e.g., Pex3p) are routed to peroxisomes via the ER. A retrograde peroxisome-to-ER transport might also exist. A novel vesicular mitochondria-to-peroxisome trafficking route has been described. A well defined sequence of morphological changes of peroxisomes, including elongation (growth), constriction, and final fission (division) contributes to peroxisome proliferation in mammalian cells. Pex11βp is involved in the elongation (tubulation) of peroxisomes, whereas DLP1, Mff and Fis1 mediate peroxisomal (and mitochondrial) fission. Proper intracellular distribution of peroxisomes formed by fission requires microtubules (MT) and motor proteins. Excess organelles are degraded by autophagy (pexophagy, mitophagy).

(adapted from Camoes et al., 2009).

1.2.2

The division machinery

Mammalian peroxisomes display a remarkably dynamic morphology. While they may appear as spherical organelles within the cytoplasm, they form tubular structures that acquire a “bead-on-a-string”-like morphology prior to their fragmentation into smaller organelles (Schrader et al., 1996a; Schrader et al., 1998b; Koch et al., 2003; Koch et al., 2004). This sequence of events is indicative of peroxisomal growth and division and occurs in a multi-step fashion by the action of a set of evolutionary conserved proteins throughout the yeast, mammalian and plant systems. Initial elongation of the peroxisomal membrane is mediated by the Pex11 family of proteins, and after subsequent constriction by a yet unidentified mechanism, final fission is carried out by dynamin-like GTPases (such as mammalian DLP1) that are recruited to the peroxisomal membrane by distinct membrane adaptors (Fis1, Mff)

(Fig. 1.5). Additionally, peroxisomes may also interconnect to form tubulo-reticular networks and a variety of morphologically distinct types of peroxisomes have been observed in different organs of mammalian organisms and cell lines (Hicks & Fahimi, 1977; Gorgas, 1987; Yamamoto & Fahimi, 1987; Roels et al., 1991; Fahimi et al., 1993; Schrader et al., 1994; Litwin & Bilinska, 1995; Schrader et al., 1996a; Schrader et al., 2000). In addition to growth and division, more complex structures such as elongated tubules or a peroxisomal reticulum may be related to other peroxisomal processes (e.g in metabolism, membrane signalling or stress protection), but information on the exact correlation between peroxisome dynamics/morphology and function is scarce.

1.2.2.1 The Pex11 family of proteins

The members of the Pex11 family of proteins represent a number of peroxisomal membrane proteins in fungi, plants and mammals that control peroxisome proliferation and regulate peroxisome morphology, size and number (Erdmann & Blobel, 1995; Marshall et al., 1995; Abe & Fujiki, 1998; Abe et al., 1998; Schrader et al., 1998b; Lingard & Trelease, 2006). They are conserved in yeasts, plants and mammals and several proteins and/or isoforms have been identified in each kingdom (Schrader & Fahimi, 2006; Hettema & Motley, 2009; Hu, 2010; Schrader et al., 2011). While Pex11 proteins were identified based on their capacity to modulate peroxisomal membrane elongation and peroxisome abundance, it has to be noted that not all Pex11 isoforms in a given species promote peroxisome membrane elongation or proliferation. In line with this, membrane association and topology may vary across organisms, ranging from a peripheral association in S. cerevisiae to multi-membrane spanning proteins in plants and mammals for (overview, see Schrader et al., 2011). The following section will provide an overview of the different Pex11 proteins across species and will give an insight into their mechanism of action.

Initially identified in C. boidinii as a inducible membrane protein upon peroxisome proliferation (and termed PMP31/32), the functional significance of Pex11p was only recognized upon deletion of its homologue in the yeast S. cerevisiae (ScPex11p) which led to the formation of one or two giant peroxisomes (Goodman et al., 1986; McCammon et al., 1990; Moreno et al., 1994; Erdmann & Blobel, 1995; Marshall et al., 1995; Sakai et al., 1995; Ma et al., 2006)} Besides Pex11p, additional proteins with weak similarity to the latter that

in Trypanosoma brucei (Maier et al., 2001; Smith et al., 2002; Rottensteiner et al., 2003a; Tam et al., 2003; Voncken et al., 2003; Huber et al., 2011; Saraya et al., 2011). Moreover,

ScPex28p/Pex29p, ScPex30p, ScPex31p/Pex32p as well as YlPex23p/Pex24p were shown to

modulate peroxisome number (Brown et al., 2000; Tam & Rachubinski, 2002; Vizeacoumar et al., 2003). Other yeast species such as H. polymorpha contain Pex11C that shares a higher similarity with ScPex11p (Kiel et al., 2006), resembling the situation in mammals. Similarly, filamentous fungi also express three Pex11 isoforms that are involved in peroxisome proliferation and Woronin body differentiation (Kabeya et al., 2005; Kiel et al., 2005b; Kiel et al., 2006; Escano et al., 2009). Plants possess five obvious homologues of Pex11p (Pex11a-e) which display differences in their expression pattern and some functional redundancy (Lingard & Trelease, 2006; Orth et al., 2007; Hu, 2010). In the mammalian system, three Pex11p isoforms were identified that control peroxisome proliferation under both basal and induced conditions: Pex11pα, Pex11pβ and Pex11pγ (Abe & Fujiki, 1998; Abe et al., 1998; Passreiter et al., 1998; Schrader et al., 1998b; Li et al., 2002a; Tanaka et al., 2003; Shimizu et al., 2004). While Pex11pβ is constitutively expressed in all tissues, both Pex11pα and Pex11pγ display tissue-specific expression patterns, but are most prominent in the liver (Passreiter et al., 1998; Schrader et al., 1998b; Li et al., 2002a; Li et al., 2002b; Tanaka et al., 2003). Among the three isoforms, only Pex11pα is induced by peroxisome proliferators activating the nuclear transcription factor PPARα (1.2.4) and is thus regarded as the regulatable mammalian Pex11p isoform (Shimizu et al., 2004). Nonetheless, Pex11pα was shown to be dispensable for PPARα-mediated peroxisome proliferation in Pex11α knock-out mice and only required to mediate peroxisome proliferation after treatment with non-classical PPARα independent proliferators (Li et al., 2002a). Furthermore, the Pex11α knock-out (KO) mouse is viable and shows no obvious effects on peroxisome number or metabolism (Li et al., 2002a), Pex11β KO, however, causes neonatal lethality and defects similar to the ZS phenotype (Li et al., 2002b), confirming its role as the central regulator of peroxisome proliferation in mammals. As expected, peroxisome abundance in Pex11β KO mice is reduced, but peroxisomal protein import and metabolism are only slightly affected. A recent comparative analysis of primary neuronal cultures and brain samples from wild-type mice, Pex11β homozygous and heterozygous knock-outs indicated a higher degree of cell death in heterozygous than in wild-type mice (Ahlemeyer et al., 2012). Moreover, heterozygotes also showed delayed neuronal differentiation, indicating that deletion of a single allele of Pex11β already causes neuronal defects in mice, a factor underappreciated so far. Thus, as dysfunctions in PEX11 do not result in any large scale alterations of peroxisomal metabolism,

patients with a PEX11 defect might have remained undetected so far. All mammalian Pex11 isoforms are tightly associated with the peroxisomal membrane and are capable of forming homo-dimers (Li & Gould, 2003; Kobayashi et al., 2007; Koch et al., 2010). Hetero-dimers were also observed, however, no interaction between Pex11pα and Pex11pβ was detected (Koch et al., 2010). Furthermore, an interaction of Pex11pβ with Fis1, a tail-anchored protein involved in the recruitment of DLP1 (1.2.2.3), has been demonstrated (Kobayashi et al., 2007; Koch et al., 2010) (Fig. 1.6). Interestingly, mammalian Pex11pβ was not only shown to be able to promote peroxisome elongation upon expression, but was also observed to concentrate at constriction sites, indicating a non-uniform distribution of the protein at the peroxisomal membrane (Schrader et al., 1998b). Recently, its role as a key component in an multistep-maturation process of asymmetric peroxisomal growth and division was further characterized (Delille et al., 2010). Employing a dominant-negative Pex11pβ-mYFP (which blocked peroxisome growth and division at an early stage) and subsequent ultrastructural and pulse-chase analysis, Pex11pβ was indicated to initially localize to spherical, pre-existing organelles where initiates the formation of a nose-like protrusion at only one side of the peroxisome. The protrusion extends to form a membrane tubule that acquires a specific set of PMPs, segments and becomes import-competent for peroxisomal matrix proteins prior to its final fission by the action of Fis1 and DLP1 (Fig. 1.6). Importantly, predominantly newly-synthesized matrix proteins are imported into the newly formed constrictions, pointing to an inherent mechanism of peroxisomal quality control linked to growth and division (Delille et al., 2010). Transient expression of various Pex11 family members of different origins led to the formation of similar membrane protrusions in mammalian cells which developed into large stacks of peroxisomal membranes (Koch et al., 2010). This pattern of Pex11p-dependent formation of specific membrane subdomains and its role in inducing a differential distribution of PMPs was also detected in the yeast H. polymorpha (Cepinska et al., 2011).

The membrane deforming capacities of the various Pex11 proteins were recently linked to the presence of several N-terminal motifs within Pex11p that are conserved in yeast, fungi and human proteins and display amphipathic properties (Opalinski et al., 2011). Negatively charged liposomes, resembling the phospholipid composition of peroxisomes, were shown to hyper-tubulate upon the addition of a Pex11 peptide containing the most conspicuous amphipathic helix of P. chrysogenum. The conservation of the amphipathic properties and its

induced by the insertion of an amphipathic helix into one leaflet of the lipid bilayer which causes membrane asymmetry and bending (Drin & Antonny, 2010) (Fig. 1.6).

Fig. 1.6: Model of peroxisomal growth and division in mammalian cells.

A well defined sequence of morphological changes of peroxisomes, including elongation (growth), constriction, and final fission (division) contributes to peroxisome proliferation in mammalian cells. Targeting to and/or activation of Pex11pβ at pre-existing peroxisomes initiates membrane remodelling and the formation of a tubular membrane extension on one side of the peroxisome. (a) Peroxisomal membrane remodelling via Pex11p is induced by the insertion of an amphipathic helix into one leaflet of the lipid bilayer which causes membrane asymmetry and bending (based on data obtained with HpPex11p). Homodimerization may keep Pex11p in an inactive form. Subsequently, the extension grows and acquires a specific set of PMPs (e.g. Pex11pβ, Fis1), before it constricts and starts to import predominantly newly-synthesized matrix proteins. Pex11pβ and the Mff-DLP1 complex concentrate at the sites of con-striction, possibly driven by alterations in membrane curvature. (b) Cytosolic DLP1 is recruited by the membrane receptor Mff. After targeting, DLP1 self-assembles into large ring-like structures that hydrolyze GTP and sever the peroxisomal membrane. Fis1 may fulfil a regulatory function. (from Schrader et al., 2011).

In regard to the regulation of Pex11p itself by post-translational modifications and/or mechanisms, monomeric ScPex11p was suggested to be inactivated by homo-dimerization, hence dimerization was proposed to regulate membrane remodelling (Marshall et al., 1996). Furthermore, phosphorylation of ScPex11p at a S165/167 residue was recently shown to be required for Pex11p action (Knoblach & Rachubinski, 2010). However, studies on the post-translational regulation of Pex11 protein activity remain restricted to the yeast system up until now.

1.2.2.2 Peroxisome fission

The final fission step of peroxisomal growth and division is executed by the action of dynamin-like mechano-enzymes, i.e. the proteins dnm1 (and vps1) in yeast species,

Koch et al., 2003; Li & Gould, 2003; Mano et al., 2004; Kuravi et al., 2006; Fujimoto et al., 2009; Kaur & Hu, 2009; Zhang & Hu, 2010). The dynamin superfamily of large GTPases includes the classical dynamins, dynamin-like proteins, Mx proteins and mitofusins in eukaryotic cells. They facilitate budding and scission events of transport vesicles, cytokinesis as well as the organelle division and fusion (for review, see Praefcke & McMahon, 2004; Heymann & Hinshaw, 2009). Classical dynamins contain five characteristic domains: the highly conserved GTPase domain as well as a middle domain and a GTP effector domain which are involved in the oligomerization and subsequent stimulation of GTPase activity. Additionally, they possess a PH domain facilitating lipid binding and a PRD mediating protein-protein interactions. Dynamin-like proteins lack one or more of the five classical domains and/or have acquired additional ones. Dynamins act as mechano-enzymes that constrict and deform membranes upon GTP hydrolysis (Sweitzer & Hinshaw, 1998; Takei et al., 1998; Takei et al., 1999). Mammalian dynamin-like protein 1 (abbreviated as DLP1 in the following) was localized to a variety of organelles, including the perinuclear region and the ER (Imoto et al., 1998; Yoon et al., 1998), but is primarily involved in the division of both mitochondria and peroxisomes (Bleazard et al., 1999; Labrousse et al., 1999; Sesaki & Jensen, 1999; Yoon et al., 2001; Koch et al., 2003; Li & Gould, 2003; Yoon, 2004). Furthermore, it contributes to the sorting of GPI-anchored, apical transport carriers at the Golgi complex (Bonekamp et al., 2010). Notably, DLP1 was the first protein identified to be a shared component of peroxisomal and mitochondrial fission (Schrader, 2006; Delille et al., 2009)(Fig. 1.7). The concept of sharing fission components between the two organelles extends to other kingdoms, as e.g. dnm1 in S. cerevisiae and DRP3B in A. thaliana are also implied in both peroxisomal and mitochondrial division (Kuravi et al., 2006; Fujimoto et al., 2009). Interestingly, the yeast S. cerevisiae uses two different types of DLPs, vps1 and dnm1, to mediate peroxisome fission. Both can complement each other, however, vps1 is utilized on glucose-grown conditions, while peroxisome proliferation induced upon growth on oleate depends on dnm1 (Motley & Hettema, 2007). At the peroxisomal membrane, mammalian DLP1 aligns in a spot-like pattern around the elongated, constricted tubules (Koch et al., 2003; Li & Gould, 2003; Koch et al., 2004); in this respect, membrane constriction might facilitate DLP1 association. DLP1 silencing by RNAi leads to the generation of elongated, constricted peroxisomes that display a prominent “beads-on-a–string”-like morphology,